My changes to the euler order in #1595 probably messed things up quite bad for some people. I originally thought I messed things up big time, but I think it's mainly just the change to the euler order, where I made it so roll no longer affected yaw estimation, but pitch did... I'm guessing people just have their trackers tilted too far forward/backward.

Manual mounting orientation being applied after attachmentFix may also be a source of problems, since a change in one variable will affect the other.

tl;dr: This will revert to the original behaviour but with acceleration fixed.